The Sampling Scheme Matters: Pan Troglodytes Troglodytes and P

Home , Central chimpanzee, Eastern chimpanzee

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 156:181–191 (2015)

The Sampling Scheme Matters: Pan troglodytes troglodytes and P. t. schweinfurthii Are Characterized by Clinal Genetic Variation Rather Than a Strong Subspecies Break

Tillmann Funfst€ uck,€ 1* Mimi Arandjelovic,1 David B. Morgan,2 Crickette Sanz,3 Patricia Reed,2 Sarah H. Olson,2 Ken Cameron,2 Alain Ondzie,4 Martine Peeters,5 and Linda Vigilant1*

1Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, 04103 Leipzig, Germany 2Wildlife Conservation Society, Global Conservation, 2300 Southern Boulevard, New York, NY 10460, USA 3Department of Anthropology, Washington University, 1 Brookings Drive, Saint Louis, MO 63130, USA 4Wildlife Conservation Society, Congo Program, 151 Ave Charles de Gaulle, Brazzaville, Republic of Congo 5UMI233, TransVIHMI, Institut de Recherche pour le Developpement (IRD) and Universite Montpellier 1, Montpel- lier, France

KEY WORDS genetic differentiation; isolation by distance; structure; microsatellites; genotyping

ABSTRACT Populations of an organism living in were sampled across large parts of their range and analyzed marked geographical or evolutionary isolation from other them together with 283 published eastern chimpanzee geno- populations of the same species are often termed subspecies types from known localities. We observed a clear signal of and expected to show some degree of genetic distinctiveness. isolation by distance across both subspecies. Further, we Thecommonchimpanzee(Pan troglodytes) is currently found that a large proportionofcomparisonsbetween described as four geographically delimited subspecies: the groups taken from the same subspecies showed higher western (P. t. verus), the nigerian-cameroonian (P. t. ellioti), genetic differentiation than the least differentiated the central (P. t. troglodytes) and the eastern (P. t. schwein- between-subspecies comparison. This proportion decreased furthii) chimpanzees. Although these taxa would be substantially when we simulated a more clumped sampling expected to be reciprocally monophyletic, studies have not scheme by including fewer groups. Our results support the always consistently resolved the central and eastern chim- general concept that the distribution of the sampled individ- panzee taxa. Most studies, however, used data from individuals can dramatically affect the inference of genetic popula- uals of unknown or approximate geographic provenance. tion structure. With regard to chimpanzees, our results Thus, genetic data from samples of known origin may shed emphasize the close relationship of equatorial chimpanzees light on the evolutionary relationship of these subspecies. from central and eastern equatorial Africa and the difﬁcult We generated microsatellite genotypes from noninvasively nature of subspecies deﬁnitions. Am J Phys Anthropol collected fecal samples of 185 central chimpanzees that 156:181–191, 2015. VC 2014 Wiley Periodicals, Inc.

INTRODUCTION Additional Supporting Information may be found in the online Understanding the distribution of genetic variation version of this article. within an endangered species such as the common chimpanzee, Pan troglodytes, is crucial for elucidating its evo- Abbreviations: HWE, Hardy-Weinberg equilibrium; IBD, isolation lutionary history (Morin et al., 1994; Stone et al., 2002; by distance; LD, linkage disequilibrium; MCP, minimum convex Gonder et al., 2006; Becquet et al., 2007), especially polygon; because there is an almost complete absence of a fossil Grant sponsor: Max Planck Society, the Arcus Foundation, the record (McBrearty and Jablonski, 2005). In addition, Paul G. Allen Family Foundation, the Dunemere Foundation, the knowledge of population relationships might also facili- Bradley L. Goldberg Foundation, the Societ e pour la Conservation tate the use of the limited resources available for conser- et le Developpement (SCD), the Wildlife Conservation Society vation efforts (Schonewald-Cox et al., 1983; Avise, 1996) (WCS).; Grant sponsor: National Institutes of Health; Grant num- and help in guiding breeding programs of chimpanzees ber: R01 AI50529.; Grant sponsor: Agence Nationale de Recherches kept in captivity (reviewed in Witzenberger and Hoch- sur le SIDA, France; Grant number: ANRS 12255. kirch, 2011; Hvilsom et al., 2013). Arguing on the basis of morphological or genetic evi- *Correspondence to: Tillmann Funfst€ uck;€ Max Planck Institute for Evolutionary Anthropology; Deutscher Platz 6; 04103 Leipzig; dence and consideration of geographic distribution, Germany. E-mail: [email protected] or Linda Vigi- researchers have separated Pan troglodytes into three lant; E-mail: [email protected] different subspecies: the western chimpanzee (P. t. verus), the central chimpanzee (P. t. troglodytes), and the Received 12 May 2014; accepted 6 October 2014 eastern chimpanzee (P. t. schweinfurthii) (e.g. Hill, 1969; Morin et al., 1994; Groves, 2001; Becquet et al., 2007). DOI: 10.1002/ajpa.22638 Later, the existence of a fourth subspecies, the Nigerian- Published online 20 October 2014 in Wiley Online Library Cameroonian chimpanzee, P. t. ellioti (Oates et al., (wileyonlinelibrary.com).

Ó 2014 WILEY PERIODICALS, INC. 182 T. FUNFST€ UCK€ ET AL.

Fig. 1. The map of Africa on the left side indicates the ranges of chimpanzee subspecies (yellow: P. t. verus, green: P. t. ellioti, red: P. t. troglodytes, blue: P. t. schweinfurthii). The zoomed in maps on the right illustrate the region sampled for the present study with blue and red dots depicting sampling locations. The red dots refer to geographically outlying samples that were not assigned to a certain group for calculations of IBD and genetic differentiation (see Methods). National borders are in red, national parks in yellow, rivers in blue and roads in white. [Color ﬁgure can be viewed in the online issue, which is available at wileyonlinelibrary. com.]

2009), originally termed P. t. vellerosus, was suggested which individuals found closer together are genetically (Gonder et al., 1997). The validity of that taxon was sup- more similar than those further apart, and it arises ported by studies utilizing mitochondrial DNA (mtDNA) when the distance an individual may disperse is smaller (Gonder et al., 1997; Gagneux et al., 1999; Gonder et al., than the continuous distribution of the organism 2006; Bjork et al., 2010), microsatellite loci (Gonder (Wright, 1943). To detect such a cline in genetic varia- et al., 2011), autosomal sequence data (Bowden et al., tion one needs to use samples of known geographic ori- 2012) as well as complete genomes (Prado-Martinez gin sampled according to a pattern that is as little et al., 2013). Thus, the currently most widely accepted clumped and as widespread as possible. However, most taxonomy of the common chimpanzee splits them into studies that detected clearly separated subspecies of Pan four geographically defined subspecies (Fig. 1), separated troglodytes used, at least partly, samples with unknown from each other from west to east by the Dahomey gap, exact origins or rather clustered sampling schemes (Bec- the Sanaga River and the Ubangi River. quet et al., 2007; Bowden et al., 2012; Prado-Martinez However, this classification is not without controversy. et al., 2013). Fischer et al. (2006) generally questioned the subspecies Although not challenging the concept of chimpanzee concept in Pan troglodytes. Using nuclear nonrepetitive subspecies in general, some authors have suggested that DNA sequences they found an extent of genetic differen- central and eastern chimpanzees do not represent geo- tiation among subspecies comparable to that seen among graphically separated subspecies, but might rather form human populations. They speculated that a more one taxon of equatorial chimpanzees. Specifically, mor- geographically-informed sampling would reveal a pat- phological studies found a close dental (Pilbrow, 2006) tern of isolation by distance (IBD) as has been described and cranial (Guy et al., 2003) similarity between P. t. for within-continental variation in humans (Serre and troglodytes and P. t. schweinfurthii. Studies describing Pa€abo,€ 2004; Rosenberg et al., 2005; Lao et al., 2008; the distribution of genetic variation within chimpanzees Tishkoff et al., 2009). IBD describes the situation in could potentially evaluate the evolutionary independence

American Journal of Physical Anthropology STRUCTURE OF EQUATORIAL CHIMPANZEES 183 of these taxa by investigating if monophyletic clades of For the present study noninvasively collected fecal central and eastern chimpanzees are found. Gagneux samples of 185 central and published genotypes from et al. (1999), analyzing mtDNA sequences of 154 eastern 283 eastern chimpanzees of known geographic origin and 24 central chimpanzees, found the eastern clade to were used. The origins of the individuals were as contin- be nested within the central clade, thereby raising the uously distributed as was feasible and included central question if these two subspecies should not be regarded and eastern chimpanzees from their western- and as one clade of equatorial chimpanzees. Later studies eastern-most range, respectively. We generated geno- including more mtDNA haplotypes again did not find types at 12 microsatellite loci and analyzed them by consistent support for monophyly of P. t. troglodytes and applying multiple methods. We used a Bayesian based P. t. schweinfurthii (Gagneux et al., 2001; Gonder et al., clustering algorithm implemented in STRUCTURE as 2006). In addition, in the study by Gonder et al. (2006) well as assessments of genetic differentiation, including no fixed nucleotide differences distinguishing the haplo- an analysis of IBD, to ask whether our data support a types of central and eastern chimpanzees were detected. distinction between central and eastern chimpanzees as These findings were, however, based on the use of a in the traditional chimpanzee taxonomy or rather if single genetic locus. A more comprehensive determina- these may be better described as one taxon of equatorial tion of the phylogeographic history of a species might chimpanzees. be obtained by use of multiple autosomal loci. Gonder et al. (2011) characterized 94 chimpanzees from all four METHODS subspecies at 27 autosomal microsatellite loci. In accordance with their mtDNA results (Gonder et al., Study area and sample collection 2006) central and eastern chimpanzees were found as Fecal samples from wild central chimpanzees were one population of equatorial chimpanzees. Even though obtained within and near Nouabale-Ndoki National Park some structuring into three different subpopulations (NP), Odzala NP (both Republic of Congo), and Lobek e was visible within equatorial chimpanzees, the genetic NP (Cameroon) (Fig. 1). Samples were collected noninva- differentiation between these was low and insignificant. sively between 1999 and 2013 and stored in RNAlater Thus, they speculated that, until recently, the equato- (samples from Lobek e) or processed by short term stor- rial chimpanzees displayed a geographical gradient of age in ethanol followed by preservation with silica gel allele frequencies with ongoing gene flow between (samples from Odzala and the Nouabale-Ndoki NP) groups (Gonder et al., 2011). Fischer et al. (2011) inves- (Nsubuga et al., 2004). tigated the genetic variation in chimpanzees and bonobos using DNA sequences from 15 autosomal regions DNA extraction, quantification, and amplification totaling some 1,50,000 base pairs. Central and eastern chimpanzees did not form monophyletic groups for any We extracted DNA from fecal samples using the of the 15 loci. QIAamp Stool kit (QIAGEN) following manufacturer’s Multiple studies (Becquet et al., 2007; Fischer et al., instructions with slight modifications (Nsubuga et al., 2011; Gonder et al., 2011; Bowden et al., 2012) that 2004). To estimate the amount of amplifiable DNA in the aimed at determining the genetic structure of chimpan- extracts we used a real-time quantitative PCR with Max- zees have made use of Bayesian based clustering algo- ima SYBRVR Green (Thermo Scientific) as Master Mix and rithms implemented in software such as STRUCTURE c-myc_E3_F1U1 plus c-myc_E3_R1U1 (Morin et al., 2001) (Pritchard et al., 2000). This approach has the advant- as primers to amplify a portion of the c-myc gene. Stand- age of not requiring a priori assignment of individuals ard curves were constructed from serially diluted human into delineated populations. However, datasets should be genomic DNA (BIO-35025, Bioline). Initially, three inde- tested for an underlying pattern of IBD before relying pendent amplifications from each extract were performed on its results, because there is considerable evidence at 12 microsatellite loci (D3s2459, D3s3038, D4s1627, that IBD alone or in combination with a clumped sam- D5s1470, D6s1056, D7s817, D7s2204, D10s676, D11s2002, pling pattern can lead clustering algorithms to overesti- D12s66, D14s306, D18s536), using a two-step multiplex mate the number of populations. In simulations PCR (Arandjelovic et al., 2009). For sex determination we conducted by Schwartz and McKelvey (2009), STRUC- also amplified a segment of the X–Y homologous ameloge- TURE correctly inferred the existence of only one popu- nin gene in a one-step PCR (Bradley et al., 2001). lation in datasets with gradual genetic variation when We also used 30 DNA extracts from chimpanzee fecal the samples were randomly distributed. However, when samples collected in a similar manner in Loango they modeled a clumped sampling scheme they accord- National Park, Gabon (Boesch et al., 2007). These ingly detected multiple clusters using STRUCTURE. By extracts were previously genotyped at D3s3038, using subsamples of an extensive empirical dataset from D5s1470, D6s1056, D10s676, D11s2002, D14s306 and Rosenberg et al. (2002), Serre and Pa€abo€ (2004) revealed typed for sex (Arandjelovic et al., 2011). We performed that in humans, whose genetic variation exhibits typi- the second step of the two-step multiplex PCR (Arandje- cally clinal variation in allele frequencies (Cavalli-Sforza lovic et al., 2009) to further genotype them at D3s2459, et al., 1994), clusters are less apparent when the sam- D4s1627, D7s817, D7s2204, D12s66, D18s536. pling is more evenly distributed geographically (but see The ABI PRISM 3100 Genetic Analyzer was used for also Rosenberg et al., 2005). Moreover, Frantz et al. electrophoresis of the PCR products. The sizes of the (2009), showed that even with a random sampling pat- alleles relative to an internal size standard were deter- tern all Bayesian programs (including STRUCTURE) mined using GeneMapper Software version 3.7 (Applied detected more than one cluster in datasets that did not Biosystems). We calculated allelic dropout rates to assess contain any genetic discontinuities, but did contain high how many observations from independent reactions are levels of IBD. Thus, results obtained by STRUCTURE necessary to confirm homozygosity (Morin et al., 2001). reported in studies that also found IBD or did not test Heterozygous genotypes were always confirmed with for IBD should be treated with caution. 99% certainty by observing each allele two or more times

American Journal of Physical Anthropology 184 T. FUNFST€ UCK€ ET AL.

TABLE 1. Overview about groups between which genetic differentiation measured as FST was determined Subspecies Country Area Group/Community N Central chimpanzees Gabon Loango NP Loango 30 Cameroon Lobek e NP Lobek e-West 12 Lobek e-East 8 Congo Odzala NP Odzala-West 8 Odzala-East 5 Nouabale Ndoki NP A 25 B1 35 C31 D21 Eastern chimpanzees Uganda Budongo Forest Reserve Busingiro 41 Kasokwa 5 Sonso 29 Kibale NP Kanyantale 49 Kanyawara 40 Ngogo 119 Totals 4 countries 6 areas 15 groups 458 individuals

chimpanzee samples as in Arandjelovic et al. (2010). After attributing samples to the correct species, we added those samples to our dataset that were genetically identiﬁed as chimpanzees and removed samples that were genetically identiﬁed as gorillas.

Data analysis In addition to the central chimpanzees that were either completely (individuals from Republic of Congo and Cameroon) or partly (individuals from Loango, Gabon) genotyped for the present study, we made use of previously published genotypes of 283 eastern chimpanzees from six communities (Langergraber et al., 2009; Langergraber et al., 2011) (Table 1 and Fig. 2). Using CERVUS 3.0 (Kalinowski et al., 2007) we analyzed the dataset consisting of 185 central and 283 eastern chimpanzees to check for the existence of null alleles within Fig. 2. Location of chimpanzee groups (black dots) between four geographically distinct subsets: (i) Loango central which genetic differentiation was measured. Countries in which these groups can be found are indicated. Cross hatch depicts chimpanzees, (ii) north-eastern central chimpanzees, (iii) central and simple hatch depicts eastern chimpanzee range. 1: Kibale eastern chimpanzees, and (iv) Budongo eastern Loango; 2: Odzala-West; 3: Odzala-East; 4: Lobek e-West; 5: chimpanzees. GENEPOP 4.2 (Raymond and Rousset, Lobek e-East; 6–9: Nouabale-Ndoki NP groups (A, B1, C, D); 10– 1995) was used to perform the exact tests of Guo and 12: Kibale communities (Kanyantale, Kanyawara, Ngogo); 13– Thompson (1992) for deviations from Hardy-Weinberg 15: Budongo communities (Busingiro, Kasokwa, Sonso). equilibrium (HWE) as well as the test of composite linkage disequilibrium (LD) (Weir, 1996) within the data (Taberlet et al., 1996). For each extract up to six inde- subsets. We found null allele frequency estimates >0.1 pendent reactions were analyzed per locus. for D7s2204 within the Budongo and the north-eastern central chimpanzee subsets. We thus tested for HWE Discrimination of individuals and distinguishing within each subset both before and after removal of gorilla and chimpanzee samples D7s2204. The Loango, Budongo, and Kibale subsets did not deviate significantly from HWE either before or after We used CERVUS 3.0 (Kalinowski et al., 2007) to cal- removal of D7s2204 (P > 0.05 after Bonferroni correction culate PIDsib (Waits et al., 2001), the probability that for multiple tests). The north-eastern central chimpan- samples with matching genotypes come from siblings zee subset, however, was only in HWE after exclusion of rather than from the same individual. Matching geno- D7s2204. Two pairs of loci (D3s3038 1 D10s676 and types were combined into a consensus genotype when D6s1056 1 D4s1627) were in LD (P < 0.05 after Bonfer- the PIDsib was <0.01. In the rare cases where two geno- roni correction for multiple tests) across all data subsets. types matched but the PIDsib was >0.01 the less com- True LD, however, can only occur between loci that are plete genotype was removed from further analyses. at the same chromosome. We thus conducted all down- In the field, fecal samples from gorillas are occasion- stream analyses excluding D7s2204 and using the ally mistakenly identified as chimpanzee feces, and vice remaining eleven loci. versa. We therefore checked for incorrectly identified We used STRUCTURE 2.3.3 (Pritchard et al., 2000) to samples by genotyping 10 samples of known gorilla ori- infer subdivision of populations. In STRUCTURE we gin at the 12 microsatellite loci typically analyzed in used the admixture model, correlated allele frequencies, chimpanzees (see above) and performing STRUCTURE a burn-in period of 100,000 steps and then 500,000 steps analysis of them in combination with the purported of data collection. We varied the number of clusters K

American Journal of Physical Anthropology STRUCTURE OF EQUATORIAL CHIMPANZEES 185 from one to eight and conducted 10 independent itera- subspecies. Therefore we investigated if there are tions at each K. We present the results of those runs within-subspecies pairs of groups that show higher that had the highest estimated logarithm of the proba- genetic differentiation than between-subspecies pairs of bility of the data at each K. We applied different meth- groups. To further elucidate how varying sampling ods to determine the optimum value for K. First, we schemes could lead to varying conclusions about the used Harvester Web v0.6.93 (Earl and vonHoldt, 2011) genetic structure of a population and to test whether to calculate DK (Evanno et al., 2005), a measure of previous reports of the subdivision of equatorial chim- second-order rate of change in the likelihood of K. Sec- panzees into central and eastern chimpanzees (e.g., Hill, ond, we looked for the K at which the likelihood distribu- 1969; Morin et al., 1994; Bjork et al., 2010; Prado- tion began to plateau (Pritchard et al., 2000) and third Martinez et al., 2013) might have been biased by a we investigated for which K value K 1 1 no longer clumped sampling scheme, we simulated an increasingly refined the clusters (i.e. at K 1 1 the clusters distin- clumped sampling and investigated if that leads to a guished at K were no longer split according to geogra- decreasing number of pairs that exhibit a higher genetic phy) (Tishkoff et al., 2009). differentiation within than between-subspecies. We To measure the genetic differentiation between differ- began by determining genetic differentiation between (i) ent groupings of individuals, a priori delineation of all nine central chimpanzee and six eastern chimpanzee groups is required. For the majority of central chimpan- groups (the least clumped scheme possible with our zees we did not know the community to which they available data) and then successively decreased the belonged. Because chimpanzee communities occupy lim- number of included groups to obtain an increasingly ited territories, we defined nine groupings of spatially clumped pattern: (ii) eight central chimpanzee groups clumped individuals within the central chimpanzees: (excluding Loango) and all six eastern chimpanzee Loango, Odzala-West, Odzala-East, Lobek e-West, groups; (iii) six central chimpanzee groups (excluding Lobek e-East, and four from the Nouabale Ndoki Loango, Odzala-West and Odzala-East) and all six east- National Park (A, B1, C, D), excluding seven geographi- ern chimpanzee groups; (iv) six central chimpanzee cally outlying samples from Odzala and three from groups (excluding Loango, Odzala-West and Odzala- Lobek e. Each grouping represents a geographic grouping East) and three eastern chimpanzee groups (excluding which is separated from the other groupings either by the three Budongo communities). distance (Loango, Lobek e-West, Lobek e-East) or by watercourses running between them (A, B1, C, D) (Figs. RESULTS 1 and 2). For the eastern chimpanzees community mem- Genotype analysis bership was known and we accordingly considered each of the six eastern chimpanzee communities to represent We genotyped 278 putative central chimpanzee fecal a distinct group. For each of the central chimpanzee samples at 12 autosomal microsatellite loci. These sam- groupings the sizes of their minimum convex polygons ples yielded 242 usable genotypes representing 158 (MCPs) were determined using ESRIVR ArcMapTM 9.2. unique individuals. An analysis of these genotypes using They ranged from 5.1 km2 (Loango) to 64.7 km2 (C), STRUCTURE in comparison with known gorilla DNAs which are broadly similar to the estimated territories of typed at the same loci showed that 15 of the purported known eastern chimpanzee communities with sizes from chimpanzee DNAs were actually from gorillas, leaving a 7km2 at the Sonso community (Newton-Fisher, 2003) total of 143 different chimpanzees. Field researchers up to 35 km2 at the Ngogo community (Mitani et al., thus correctly identified chimpanzee dung in 227/ 2010). A total of 15 chimpanzee groups (nine central and 242 5 94% of cases. In a reciprocal analysis of genotypes six eastern) were defined (Table 1 and Fig. 2). Using purportedly from gorillas and typed as part of a comple- Arlequin 3.5. (Excoffier and Lischer, 2010) we calculated mentary study (Funfst€ uck€ et al., 2014), we found 16 the FST values between all 105 possible pairs of groups samples were derived from chimpanzees. These 16 sam- and correlated them with the corresponding geographic ples were secondarily genotyped at microsatellite loci distances (measured as natural logarithm of distance in used here. Four samples were identical to already exist- meters). For the calculations of the geographic distances, ing genotypes and 12 represented additional individuals. we assessed the positions of the MCP centroids of the The final dataset therefore contained 155 different cen- central chimpanzee groupings using ESRIVR ArcMapTM tral chimpanzee genotypes, which were on average 88% 9.2. Regarding the eastern chimpanzee groups we complete (Supporting Information Table 1). These 155 obtained the positions of the MCP centroids for Ngogo individuals consisted of 74 females (F) and 78 males (M) and Kanyantale (S. Amsler, personal communication), while the sex could not be determined for three individu- whereas central points within each territory were used als. Twenty individuals were from Odzala (13F, 7M), 23 for Busingiro, Sonso, Kasokwa, and Kanyawara (J. from Lobek e (10F, 10M, 3 unknown sex), and 112 (51F, Moore, personal communication). We then plotted a 61M) from Nouabale-Ndoki. regression line and conducted a Mantel test in R (Devel- The average DNA concentration of the 225 quantified opment Core Team, 2012) and considered its slope signif- extracts was 237.4 pg/ml (median: 108.0 pg/ml; SD: 573.4 icantly positive if there was a less than 5% probability of pg/ml; range: 3.3–6063 pg/ml). Considering results from obtaining a slope greater than the observed one in individual PCRs, allelic dropout rates over all loci were 10,000 permutations. This allowed us to see if the varia- 0.31 for extracts with the lowest DNA amounts (5 pg/ tion in genetic distance was consistent with a pattern of ml), 0.24 for extracts with intermediate DNA amounts IBD encompassing both subspecies. (5–10 pg/ml) and 0.04 for extracts with the highest DNA If the classification of central and eastern chimpanzees amounts (>10 pg/ml). Thus, four independent observa- as distinct subspecies reflects long-term, independent tions of homozygosity were needed to confirm homozy- evolutionary histories of these two taxa, one predicts gous genotypes with 99% certainty in samples with DNA that pairwise comparisons between-subspecies should concentrations 5 pg/ml (0.31^4 5 0.009), whereas three show a higher genetic differentiation than those within- independent observations were sufficient for all other

American Journal of Physical Anthropology 186 T. FUNFST€ UCK€ ET AL.

Fig. 3. STRUCTURE analysis including central chimpanzees from the north-eastern region (present study) (bars 1–155) and from Loango, Gabon (Arandjelovic et al., 2011 and present study) (bars 156–185) as well as eastern chimpanzees from Budongo, Uganda (Langergraber et al., 2011) (bars 186–260) and from Kibale NP, Uganda (Langergraber et al., 2009) (bars 261–468). Colors represent the inferred ancestry from K ancestral populations. Ten runs were conducted at each K. Shown are the results for K 5 2 to K 5 5, presenting those runs that had the highest estimated logarithm of the probability of the data. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.] extracts (0.24^3 5 0.01 for 5–10 pg/ml and Genetic differentiation 0.04^3 5 0.00006 for >10 pg/ml). We also genotyped 30 extracts that were known to While the results of our STRUCTURE analysis seem originate from different central chimpanzees (Arandje- to indicate that equatorial chimpanzees can be divided lovic et al., 2011) at six additional autosomal microsatel- into at least two demes that seemingly correspond to lite loci. These six-loci genotypes were 72.2% complete. central and eastern “subspecies,” pairwise comparisons However, the genotypes of the other six-loci done by Ara- between the 15 groups of both putative subspecies ndjelovic et al. (2011) were 99.4% complete leading to showed that genetic differentiation increased signifi- 85.8% complete genotypes over all 12 loci (Supporting cantly with geographic distance, a pattern suggestive of Information Table 1). strong IBD effects (Fig. 4). The FST values of pairwise We conducted analyses using STRUCTURE to exam- comparisons ranged from slightly negative values, sug- ine the distribution of individuals into clusters. Different gesting no differentiation, to 0.1456 between Lobek e- methods of inferring the optimal value of K produced dif- West and Kasokwa (Table 2). The average FST values ferent inferences. The first method (using DK) supported between two groups of central chimpanzees, two groups K 5 2, the second one (using the K at which the likeli- of eastern chimpanzees and a central versus an eastern hood distribution begun to plateau) produced ambiguous chimpanzee group were 0.0164, 0.0426 and 0.0802, results but suggested K 5 3 and the third one (using the respectively. When we used all nine central and six east- K at which K11 no longer refined the clusters) found ern chimpanzee groups for calculations of genetic differ- K 5 4. The STRUCTURE results for K 5 2toK 5 5 are entiation, we found that 54.90% (28/51) of all within- shown (Fig. 3). At K 5 2 the analysis of 185 central and subspecies comparisons exhibited higher values of FST 283 eastern chimpanzees produced distinct clusters for than the least differentiated between-subspecies compar- central and eastern chimpanzees. At K 5 3 the central ison (Table 2, Fig. 5). When we excluded the geographi- chimpanzees formed one cluster and the eastern chim- cally outlying Loango central chimpanzee sample from panzees were further subdivided into individuals from the analysis, this value changed only very little to Budongo and Kibale, whereas at K 5 4 the central chim- 55.81% (24/43). To further investigate the effects of panzees were divided into a Loango and a north-eastern increasingly limited sampling, we next excluded Odzala- cluster and the eastern chimpanzees were divided into West and Odzala-East from the central chimpanzee Budongo and Kibale clusters. Finally, at K 5 5, the clus- data, yielding into a proportion of 26.67% (8/30). Finally, ters distinguished at K 5 4 were no longer split. Instead by the omission of the three Budongo eastern chimpan- the individuals from the Kibale cluster were admixed zee communities, that value dropped again to just 5.56% and found to have ancestry from two different sources. (1/18). That is, there is a clear trend that successively

American Journal of Physical Anthropology STRUCTURE OF EQUATORIAL CHIMPANZEES 187 *** . Odzala-West; 5 not signiﬁcant, 5 Odzala-East; O-W 5 * *** *** *** ***

Fig. 4. Genetic distances measured as Fst plotted against Ngogo; O-E the natural logarithm of straight line distance in meters. Each 5 point represents one pairwise comparison between the ﬁfteen *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** groups. The linear equation for the regression line was *** *** *** *** *** *** 0.04570.0455 0.0747 0.0568 0.0630 0.0609 0.0018 0.0155 0.0117 *** y 5 0.0146x 2 0.1358 with r 5 0.651 and P 5 0.0001. 0.03210.02210.0439 0.0538 0.0539 0.0580 ** *** *** n.s *** *** *** ** e-West; Ngo . *** *** *** *** *** *** . *** ** *** ** ** ** . *** *** * *** *** ** omitting more geographically distant populations from ek *** n.s our sample results in fewer within-subspecies compari- n.s Lob

sons exhibiting higher FST values than the least differ- 5 entiated between-subspecies pair, therefore reiterating . . the importance of IBD in explaining the genetic differen- . **** **** ** *** ** *** n.s tiation among central and eastern equatorial populations n.s of chimpanzees. e-East; L-W

DISCUSSION ek ...... Lob

In this study we investigated the population structure 5 n.s of central and eastern chimpanzees by subjecting the 0.01990.0050 0.0168 n.s 2 2 microsatellite genotypes of 185 central and 283 eastern chimpanzees of known geographic origin to multiple analyses. First, we applied a Bayesian based clustering . algorithm implemented in STRUCTURE. Second, we Loango; L-E 0.0090 n.s 5 0.001) between nine central and six eastern chimpanzee groups tested for IBD using correlations between Euclidean and 2 <

genetic distances. Finally, we compared the genetic dif- P ferentiation between groups drawn from the same subspecies and between groups drawn from different * *** n.s subspecies. Such use of various methods is highly recom- *** *** n.s 0.0232 n.s 0.0164 0.0012 0.0019 Kasokwa; Lo 0.01, *** 2 2 mended for attempting to obtain an unbiased, accurate 2 5 <

and comprehensive view about the genetic structure of P the populations under study (Frantz et al., 2009; ) of pairwise comparisons (below diagonal) and associated levels of signiﬁcance (above diagonal, with n.s. . . Schwartz and McKelvey, 2009; Funfst€ uck€ et al., 2014). ST F * ** *** n.s n.s 0.0016

In STRUCTURE, each method of inferring the optimal 0.05, ** 2 value for K suggested different amounts of clusters. < P

Determining DK suggested two populations, consistent * with a division of the sample into eastern and central Kanyawara; Kas . . chimpanzees. Contrary to that, the likelihood distribu- 5 n.s tion began to plateau at K 5 3, suggesting one cluster of 0.0065 central chimpanzees and two clusters of eastern chim- 2 panzees. Finally, the clusters distinguished at K 5 4,

were no longer split at K 5 5, arguing for four clusters of .

geographically clumped individuals: (i) central chimpan- n.s 0.0009 n.s Kanyantale; Kw zees from Loango, (ii) north-eastern central chimpan- 2 zees, (iii) eastern chimpanzees from Budongo, and (iv) 5 eastern chimpanzees from Kibale (Fig. 3). Thus, the analyses conducted in STRUCTURE seem to support the A B1 C D L-E L-W O-W O-E Lo Bus Kas Son Ngo Kt Kw 0.0018 n.s 0.0451 0.0479 0.0288 0.0272 0.0051 0.0024 0.0109 0.0003 0.0070 0.0009 ** *** n.s 0.0366 0.0294 0.0212 0.0300 n.s 0.0453 0.0552 0.0499 0.0435 0.0132 0.0411 existence of two to four genetic clusters within equato- 0.0372 0.0413 0.0256 0.0290 2 2 2 rial chimpanzees. These STRUCTURE results, however, 2 Busingiro; Kt might lead to a wrong or at least biased perception of Sonso. 5 5

the population structure within equatorial chimpanzees TABLE 2. Genetic differentiation (measured as KtKwPairwise comparisons within 0.1163 subspeciesBus and 0.1187 between subspecies are in italic and 0.1196 normal font, respectively. 0.1245 0.0967 0.0992 0.1030 0.0979 0.0464 0.0577 0.0741 0.0802 0.0378 0.0293 0.0363 0.052434 0.0787 0.0593 Son B1 A C L-W O-W D L-E Lo O-E BusKas 0.0883 0.1326 0.0906 0.1206 0.0751 0.1145 0.0808 0.1356 0.0481 0.11354 0.0588 0.1456 0.0386 0.1194 0.0550 0.0973 0.0732 0.1174 for several reasons. First, the fact that different methods SonNgo 0.0861 0.1016 0.0913 0.1102 0.0721 0.0859 0.0708 0.0911 0.0481 0.0460 0.0485 0.0716 0.0204 0.0323 0.0550 0.0364 0.0684 0.0633

American Journal of Physical Anthropology 188 T. FUNFST€ UCK€ ET AL.

Fig. 5. Number of pairwise comparisons (y-axes) within different ranges of genetic differentiation measured as FST (x-axes). Black bars and white bars represent within and between subspecies comparisons, respectively. Groups included into the bar plots were as follows. Top left (A): nine central and six eastern chimpanzee groups. Top right (B): eight central (excluding Loango) and six eastern chimpanzee groups. Bottom left (C): Six central (excluding Loango, Odzala-West and Odzala-East) and six eastern chimpanzee groups. Bottom right (D): Six central (excluding Loango, Odzala-West and Odzala-East) and three eastern chimpanzee (excluding the three Budongo communities) groups. for inferring the exact amount of populations found dif- might well be that a geographically more continuous ferent solutions, makes it difficult to find definitive sampling would not reveal clearly distinct genetic clus- answers about the genetic stratification within central ters within equatorial chimpanzees. In sum, these and eastern chimpanzees. Second, due to its problems in results suggest that the geographic distance shapes the dealing with clinal genetic variation, the algorithm used pattern of genetic variation observed in our sample and by STRUCTURE is prone to overestimate the number of those geographic clusters are inconsistently inferred. populations when there is an underlying pattern of IBD, To evaluate if another approach suggests a primary especially when the sampling scheme is clumped (Frantz division of equatorial chimpanzee genetic variation into et al., 2009; Schwartz and McKelvey, 2009). By compar- two units, we calculated pairwise genetic differentiations ing geographic and genetic distances, we found highly between groups from the same subspecies and between significant evidence for IBD and our STRUCTURE anal- groups from different subspecies. If a taxonomical dis- ysis also identified several individuals that had less than tinction as different subspecies reflects a long period of 75% ancestry within one cluster meaning that they independent evolution, one would expect that within- appeared admixed and not attributable to a single clus- subspecies comparisons exhibit lower estimates of ter. Such high proportions of admixed individuals are genetic variation than between-subspecies comparisons. commonly detected by STRUCTURE if the distribution Moreover, if the previously reported distinction between of genetic variation is characterized by a pattern of IBD central and eastern chimpanzees is real, and not an arti- (Pritchard et al., 2003). Third, and probably most impor- fact of clumped sampling, the proportion of within- tant, although our sampling was as less clumped as fea- subspecies comparisons showing higher genetic differen- sible, there were still large geographical gaps between tiation than the least differentiated between-subspecies some sampling locations (Fig. 2). In the clustering solu- comparison should be independent of the sampling tions that were supported by our STRUCTURE results, scheme. Contrary to these predictions we found that a the suggested populations consisted of individuals that substantial proportion (54.90%) of all within-subspecies were geographically clumped together and the geo- comparisons exhibited higher values of FST than the graphic distances between samples from different clus- least differentiated between-subspecies comparison ters were much larger than those between samples from when all groups were included (which represents the the same cluster. We thus believe that, even if STRUC- least clumped scheme). After stepwise exclusion of TURE did not overestimate the amount of populations groups in order to simulate an increasingly clumped due to IBD, it simply depicts the lack of sampling at geo- sampling scheme, there was a clear trend of obtaining graphic gaps between the clusters. That is, the results of reduced proportions of within-subspecies comparisons STRUCTURE reflect the chosen sampling scheme. It exceeding between-subspecies comparisons with a final

American Journal of Physical Anthropology STRUCTURE OF EQUATORIAL CHIMPANZEES 189 value of only 5.56% in the most clumped scheme. Two eastern chimpanzees as one group. However, the range of main conclusions can be inferred from these results. The both subspecies together spans over more than 1500 km high proportions found in the least clumped schemes from north to south and more than 2000 km from east to argue for an incomplete division with recent admixture west (Fig. 2). Many chimpanzees live in small populations between central and eastern chimpanzees, whereas the within fragmented forest patches and the low connectivity second point demonstrates how much a certain sampling between these populations is considered as one of the lead- pattern affects the resulting estimates of genetic varia- ing threats to the species existence (Plumptre et al., 2010). tion and hence also the conclusions drawn from it. Thus, efforts to maintain and repair the interconnectedness Previous studies that found a clear distinction between of these fragments and mitigate detrimental impacts such central and eastern chimpanzees, although sometimes as inbreeding are warranted. Corridor conservation is a rela- employing extensive datasets (Becquet et al., 2007; Prado- tively new field in equatorial Africa, but given the expansion Martinez et al., 2013), were often relying on samples of of natural resource extraction industries in forestry and unknown exact origin, a small number of analyzed individ- mining sectors and the development of associated hard uals or a combination of both (Gonder et al., 1997; Bjork infrastructures (Laporte et al., 2007) it will likely be neces- et al., 2010; Becquet et al., 2007; Prado-Martinez et al., sary on behalf of conserving chimpanzees in the Congo 2013). In contrast, our study investigates the population Basin. The prospects for such strategies remain high given structure of central and eastern chimpanzees across large there still persist extensive chimpanzee populations in cen- parts of their range by analyzing microsatellite genotypes tral and eastern Africa that have as yet to be disturbed by of spatially explicit and widely distributed samples. Com- present day anthropogenic impacts (Morgan and Sanz, pared with the analysis of mtDNA sequence data, the use 2003; Hicks et al., 2014). Further, Junker et al. (2012), by of microsatellites has the advantage that the latter are relating ape presence information to environmental and representing selectively neutral multi-locus markers, human impact factors, predicted that geographical connec- which are biparentally inherited. We are therefore confi- tivity of patches within the Congo Basin of suitable habitat dent that our results represent a reliable description of the are promising, especially in the north-eastern part of the population structure of equatorial chimpanzees. We argue Republic of Congo and the eastern parts of the Democratic that instead of the traditional taxonomy with two clearly Republic of Congo. Future efforts integrating the relevance distinct subspecies of central and eastern chimpanzees, of these findings to conservation of local chimpanzee popula- the population structure of equatorial chimpanzees might tions in terms of genetic viability and dispersal would be better described by clinal genetic variation across their improve prioritization scenarios. range with ongoing gene flow and recent admixture rather than a strong distant split. That leads to an encouraging ACKNOWLEDGMENTS agreement with other studies on chimpanzee population structure that also did not find evidence for central and For permission to carry out research in Republic of eastern chimpanzees constituting clearly separated sub- Congo, the authors thank the Ministe`re de l’Economie For- species (Gagneux et al., 1999, 2001; Gonder et al., 2006, estie`re et Developpement Durable; the Ministe`re de la 2011; Fischer et al., 2011). As our sampling design also Recherche Scientifique et de l’Innovation Technologique; allowed us to determine to which extent a clumped sam- and the Ministe`re de la Sante. For permission to collect pling can bias the results, we recommend that further samples in Cameroon, the authors thank the Ministe`re de studies should aim to obtain spatially explicit samples in a la Recherche Scientifique et de l’Innovation (MINRESI) as preferably even less clumped sampling pattern. We specu- well as the Ministe`re des Forets^ et de la Faune (MINFOF). late that analysis of such datasets will demonstrate yet The authors thank the Agence Nationale des Parcs Natio- more clearly that equatorial chimpanzees are character- naux (ANPN) and the Centre National de la Recherche Sci- ized by IBD and that previously found subdivisions might entifique et Technique (CENAREST) of Gabon for have been artifacts caused by an incomplete sampling permission to conduct their research in Loango National scheme. Likewise to the situation in humans, where tradi- Park as well as the staff and the SIV team from PRESICA tionally the existence of separate races was proposed, clas- for logistical support in Cameroon. They also thank Afri- sical primate taxonomists were often obsessed with finding can Parks Network for facilitation of the work. They thank distinct subspecies (Hill, 1969). In humans, a more contin- A. Abraham for laboratory assistance, C. Stephens, and R. uous sampling revealed a pattern of IBD rather than typo- Mundry for statistical assistance. They thank two anony- logical races (Serre and Pa€abo,€ 2004; Lao et al., 2008; mous referees for helpful comments. They thank L. Raba- Tishkoff et al., 2009). Our results indicate that also in nal, L. Mackaga, E. R. Guizard, N. Tagg, B. Graw, E. other primate taxa an obvious line between distinct sub- Fairet, M. Gregoire, L. Rankin S. Kaba, M. J. Akongo, E. B. species may be difficult to draw. Ngouembe, and G. Bounga for sample collection. All work From a conservation point of view, chimpanzees play an presented herein was conducted in compliance with appro- extraordinarily important role, because they are not only priate animal care regulations and national laws and also the closest living relatives of humans, but also act as a flag- in compliance with the American Association of Physical ship species, umbrella species and environmental indicator Anthropologists Code of Ethics. species (Wrangham et al., 2008). Moreover, through their role as seed-dispersers they are important in maintaining LITERATURE CITED an intact forest ecosystem (Balcomb and Chapman, 2003). Both central and eastern chimpanzees are categorized as Arandjelovic M, Guschanski K, Schubert G, Harris TR, endangered by the IUCN (Walsh et al., 2003; Wilson et al., Thalmann O, Siedel H, Vigilant L. 2009. Two-step multiplex 2009) and their abundance has severely declined over the polymerase chain reaction improves the speed and accuracy last decades (Walsh et al., 2003; Hicks et al., 2010; Plumptre of genotyping using DNA from noninvasive and museum sam- et al., 2010). Our results, indicating a mostly contiguous pop- ples. Mol Ecol Resour 9:28–36. ulation of equatorial chimpanzees, might argue for large- Arandjelovic M, Head J, Kuhl€ H, Boesch C, Robbins MM, scale oriented conservation efforts considering central and Maisels F, Vigilant L. 2010. Effective non-invasive genetic

American Journal of Physical Anthropology 190 T. FUNFST€ UCK€ ET AL.

monitoring of multiple wild western gorilla groups. Biol Con- genetic structure and histories of chimpanzee populations. serv 143:1780–1791. Proc Natl Acad Sci 108:4766–4771. Arandjelovic M, Head J, Rabanal LI, Schubert G, Mettke E, Gonder MK, Oates JF, Disotell TR, Forstner MR., Morales JC, Boesch C, Robbins MM, Vigilant L. 2011. Non-invasive Melnick DJ. 1997. A new West African chimpanzee subspe- genetic monitoring of wild central chimpanzees. PLoS ONE 6: cies? Nature 388:337. e14761. Groves C. 2001. Primate taxonomy. Washington, DC: Smithso- Avise JC. 1996. Introduction: the scope of conservation genetics. nian Institution Press. In: Avise JC, Hamrick J, editors. Conservation genetics: case Guo SW, Thompson EA. 1992. Performing exact test of Hardy– histories from nature. New York: Chapman & Hall. p 1–9. Weinberg proportion for multiple alleles. Biometrics 48:361– Balcomb SR, Chapman CA. 2003. Bridging the gap: influence of 372. seed deposition on seedling recruitment in a primate-tree Guy F, Brunet M, Schmittbuhl M, Viriot L. 2003. New interaction. Ecol Monogr 73:625–642. approaches in hominoid taxonomy: morphometrics. Am J Becquet C, Patterson N, Stone AC, Przeworski M, Reich D. Phys Anthropol 121:198–218. 2007. Genetic structure of chimpanzee populations. PLoS Genet 3:e66. Hicks TC, Darby L, Hart J, Swinkels J, January N, Menken S. Bjork A, Liu W, Wertheim JO, Hahn BH, Worobey M. 2010. 2010. Trade in orphans and bushmeat threatens one of the Evolutionary history of chimpanzees inferred from complete Democratic Republic of the Congo’s most important popula- mitochondrial genomes. Mol Biol Evol 28:615–623. tions of eastern chimpanzees (Pan troglodytes schweinfurthii). Boesch C, Head J, Tagg N, Arandjelovic M, Vigilant L, Robbins Afr Primates 7:1–18. MM. 2007. Fatal chimpanzee attack in Loango National Park, Hicks TC, Tranquilli S, Kuehl H, Campbell G, Swinkels J, Gabon. Int J Primatol 28:1025–1034. Darby L, Boesch C, Hart J, Menken SBJ. 2014. Absence of Bowden R, MacFie TS, Myers S, Hellenthal G, Nerrienet E, evidence is not evidence of absence: discovery of a large, con- Bontrop RE, Freeman C, Donnelly P, Mundy NI. 2012. tinuous population of Pan troglodytes schweinfurthii in the Genomic tools for evolution and conservation in the chimpan- Central Uele region of northern DRC. Biol Conserv 171:107– zee: Pan troglodytes ellioti is a genetically distinct population. 113. PLoS Genet 8:e1002504. Hill WCO. 1969. The nomenclature, taxonomy, and distribution Bradley BJ, Chambers KE, Vigilant L. 2001. Accurate DNA- of chimpanzees. In: Bourne GH, editor. The chimpanzee, Vol. based sex identification of apes using non-invasive samples. 1. Basel: Karger. p 22–49. Conserv Genet 2:179–181. Hvilsom C, Frandsen P, Børsting C, Carlsen F, SalleB, Cavalli-Sforza LL, Menozzi P, Piazza A. 1994. The history and Simonsen BT, Siegismund HR. 2013. Understanding geo- geography of human genes. Princeton, NJ: Princeton Univer- graphic origins and history of admixture among chimpanzees sity Press. in European zoos, with implications for future breeding pro- Earl DA, vonHoldt BM. 2011. STRUCTURE HARVESTER: a grammes. Heredity 110:586–593. website and program for visualizing STRUCTURE output Junker J, Blake S, Boesch C, Campbell G, Toit L du, Duvall C, and implementing the Evanno method. Conserv Genet Resour Ekobo A, Etoga G, Galat-Luong A, Gamys J, Ganas-Swaray 4:359–361. J, Gatti S, Ghiurghi A, Granier N, Hart J, Head J, Herbinger Evanno G, Regnaut S, Goudet J. 2005. Detecting the number of I, Hicks TC, Huijbregts B, Imong IS, Kuempel N, Lahm S, clusters of individuals using the software STRUCTURE: a Lindsell J, Maisels F, McLennan M, Martinez L, Morgan B, simulation study. Mol Ecol 14:2611–2620. Morgan D, Mulindahabi F, Mundry R, N’Goran KP, Normand Excoffier L, Lischer HEL. 2010. Arlequin suite ver 3.5: a new E, Ntongho A, Okon DT, Petre C-A, Plumptre A, Rainey H, series of programs to perform population genetics analyses Regnaut S, Sanz C, Stokes E, Tondossama A, Tranquilli S, under Linux and Windows. Mol Ecol Resour 10:564–567. Sunderland-Groves J, Walsh P, Warren Y, Williamson EA, Fischer A, Pollack J, Thalmann O, Nickel B, Pa€abo€ S. 2006. Kuehl HS. 2012. Recent decline in suitable environmental Demographic history and genetic differentiation in apes. Curr conditions for African great apes. Divers Distrib 18:1077– Biol 16:1133–1138. 1091. Fischer A, Prufer€ K, Good JM, Halbwax M, Wiebe V, AndreC, Kalinowski ST, Taper ML, Marshall TC. 2007. Revising how the Atencia R, Mugisha L, Ptak SE, Pa€abo€ S. 2011. Bonobos fall computer program CERVUS accommodates genotyping error within the genomic variation of chimpanzees. PloS One 6: increases success in paternity assignment. Mol Ecol 16:1099– e21605. 1106. Frantz AC, Cellina S, Krier A, Schley L, Burke T. 2009. Using Langergraber K, Mitani J, Vigilant L. 2009. Kinship and social spatial Bayesian methods to determine the genetic structure bonds in female chimpanzees (Pan troglodytes). Am J Prima- of a continuously distributed population: clusters or isolation tol 71:840–851. by distance? J Appl Ecol 46:493–505. Langergraber K, Schubert G, Rowney C, Wrangham R, Funfst€ uck€ T, Arandjelovic M, Morgan DB, Sanz C, Breuer T, Zommers Z, Vigilant L. 2011. Genetic differentiation and the Stokes EJ, Reed P, Olson SH, Cameron K, Ondzie A, Peeters evolution of cooperation in chimpanzees and humans. Proc R M, Kuhl€ HS, Cipolletta C, Todd A, Masi S, Doran-Sheehy Soc B Biol Sci 278:2546–2552. DM, Bradley BJ, Vigilant L. 2014. The genetic population structure of wild western lowland gorillas (Gorilla gorilla Lao O, Lu TT, Nothnagel M, Junge O, Freitag-Wolf S, Caliebe gorilla) living in continuous rain forest. Am J Primatol 76: A, Balascakova M, Bertranpetit J, Bindoff LA, Comas D, 868–878. Holmlund G, Kouvatsi A, Macek M, Mollet I, Parson W, Palo Gagneux P, Gonder MK, Goldberg TL, Morin PA. 2001. Gene J, Ploski R, Sajantila A, Tagliabracci A, Gether U, Werge T, flow in wild chimpanzee populations: what genetic data tell Rivadeneira F, Hofman A, Uitterlinden AG, Gieger C, us about chimpanzee movement over space and time. Philos Wichmann H-E, Ruther€ A, Schreiber S, Becker C, Nurnberg€ Trans R Soc B Biol Sci 356:889–897. P, Nelson MR, Krawczak M, Kayser M. 2008. Correlation Gagneux P, Wills C, Gerloff U, Tautz D, Morin PA, Boesch C, between genetic and geographic structure in Europe. Curr Fruth B, Hohmann G, Ryder OA, Woodruff DS. 1999. Mito- Biol 18:1241–1248. chondrial sequences show diverse evolutionary histories of Laporte NT, Stabach JA, Grosch R, Lin TS, Goetz SJ. 2007. African hominoids. Proc Natl Acad Sci 96:5077–5082. Expansion of industrial logging in Central Africa. Science Gonder MK, Disotell TR, Oates JF. 2006. New genetic evidence 316:1451–1451. on the evolution of chimpanzee populations and implications McBrearty S, Jablonski NG. 2005. First fossil chimpanzee. for taxonomy. Int J Primatol 27:1103–1127. Nature 437:105–108. Gonder MK, Locatelli S, Ghobrial L, Mitchell MW, Kujawski JT, Mitani JC, Watts DP, Amsler SJ. 2010. Lethal intergroup Lankester FJ, Stewart C-B, Tishkoff SA. 2011. From the aggression leads to territorial expansion in wild chimpanzees. Cover: evidence from Cameroon reveals differences in the Curr Biol 20: R507–R508.

American Journal of Physical Anthropology STRUCTURE OF EQUATORIAL CHIMPANZEES 191

Morgan D, Sanz C. 2003. Na€ıve encounters with chimpanzees R Core Team. 2013. R: a language and environment for statisti- in the Goualougo Triangle, Republic of Congo. Int J Primatol cal computing. R Foundation for Statistical Computing, 24:369–381. Vienna, Austria. Morin PA, Chambers KE, Boesch C, Vigilant L. 2001. Quantita- Rosenberg NA, Mahajan S, Ramachandran S, Zhao C, tive polymerase chain reaction analysis of DNA from nonin- Pritchard JK, Feldman MW. 2005. Clines, clusters, and the vasive samples for accurate microsatellite genotyping of wild effect of study design on the inference of human population chimpanzees (Pan troglodytes verus). Mol Ecol 10:1835–1844. structure. PLoS Genet 1:e70. Morin PA, Moore JJ, Chakraborty R, Jin L, Goodall J, Woodruff Rosenberg NA, Pritchard JK, Weber JL, Cann, H. W., Kidd KK, DS. 1994. Kin selection, social structure, gene flow, and the Zhivotovsky LA, Feldman MW. 2002. Genetic structure of evolution of chimpanzees. Science 265:1193–1201. human populations. Science 298:2381–2385. Newton-Fisher NE. 2003. The home range of the Sonso commu- Schonewald-Cox CM, Chambers SM, MacBryde B, Thomas WL. nity of chimpanzees from the Budongo Forest, Uganda. Afr J 1983. Genetics and conservation: a reference for managing Ecol 41:150–156. wild animal and plant populations. Menlo Park, California: Nsubuga AM, Robbins MM, Roeder AD, Morin PA, Boesch C, Benjamin/Cummings Publishing Company. Vigilant L. 2004. Factors affecting the amount of genomic Schwartz MK, McKelvey KS. 2009. Why sampling scheme mat- DNA extracted from ape faeces and the identification of an ters: the effect of sampling scheme on landscape genetic improved sample storage method. Mol Ecol 13:2089–2094. results. Conserv Genet 10:441–452. Oates JF, Groves CP, Jenkins PD. 2009. The type locality of Serre D, Pa€abo,€ S. 2004. Evidence for gradients of human Pan troglodytes vellerosus (Gray, 1862), and implications for genetic diversity within and among continents. Genome Res the nomenclature of West African chimpanzees. Primates 50: 14:1679–1685. 78–80. Stone AC, Griffiths RC, Zegura SL, Hammer MF. 2002. High Pilbrow V. 2006. Population systematics of chimpanzees using levels of Y-chromosome nucleotide diversity in the genus Pan. molar morphometrics. J Hum Evol 51:646–662. Proc Natl Acad Sci 99:43–48. Plumptre A, Rose R, Nangendo G, Williamson EA, Didier K, Taberlet P, Griffin S, Goossens B, Questiau S, Manceau V, Hart F, Mulindahabi F, Hicks C, Griffin B, Ogawa H, Nixon Escaravage N, Waits LP, Bouvet J. 1996. Reliable genotyping S, Pintea L, Vosper A, McLennan M, Amsini F, McNeilage A, of samples with very low DNA quantities using PCR. Nucleic Makana JR, Kanamori M, Hernandez A, Piel A, Stewart F, Acids Res 24:3189–3194. Moore J, Zamma K, Nakamura M, Kamenya S, Idani G, Tishkoff SA, Reed FA, Friedlaender FR, Ehret C, Ranciaro A, Sakamaki T, Yoshikawa M, Greer D, Tranquilli S, Beyers R, Froment A, Hirbo JB, Awomoyi AA, Bodo J-M, Doumbo O, Hashimoto C, Furuichi T, Bennett E. 2010. Status survey and Ibrahim M, Juma AT, Kotze MJ, Lema G, Moore JH, conservation action plan for the Eastern chimpanzee (Pan Mortensen H, Nyambo TB, Omar SA, Powell K, Pretorius GS, troglodytes schweinfurthii), 2010–2020. Gland, Switzerland: Smith MW, Thera MA, Wambebe C, Weber JL, Williams SM. IUCN. p 48. 2009. The genetic structure and history of Africans and Afri- Prado-Martinez J, Sudmant PH, Kidd JM, Li H, Kelley JL, can Americans. Science 324:1035–1044. Lorente-Galdos B, Veeramah KR, Woerner AE, O’Connor TD, Santpere G, Cagan A, Theunert C, Casals F, Laayouni H, Waits LP, Luikart G, Taberlet P. 2001. Estimating the probabil- Munch K, Hobolth A, Halager AE, Malig M, Hernandez- ity of identity among genotypes in natural populations: cau- Rodriguez J, Hernando-Herraez I, Prufer€ K, Pybus M, tions and guidelines. Mol Ecol 10:249–256. Johnstone L, Lachmann M, Alkan C, Twigg D, Petit N, Baker Walsh P, Abernathy K, Bermejo M, Beyers R, De Wachter P, C, Hormozdiari F, Fernandez-Callejo M, Dabad M, Wilson Akou ME, Huijbregts B, Mambounga DI, Toham AK, ML, Stevison L, Camprubı C, Carvalho T, Ruiz-Herrera A, Kilbourn AM, Lahm SA, Latour S, Maisels F, Mbina C, Vives L, Mele M, Abello T, Kondova I, Bontrop RE, Pusey A, Mihindou Y, Obiang SN, Effa EN, Starkey MP, Telfer P, Lankester F, Kiyang JA, Bergl RA, Lonsdorf E, Myers S, Thibault M, Tutin CEG, White LJT, Wilkie DS. 2003. Cata- Ventura M, Gagneux P, Comas D, Siegismund H, Blanc J, strophic ape decline in western equatorial Africa. Nature 422: Agueda-Calpena L, Gut M, Fulton L, Tishkoff SA, Mullikin 611–614. JC, Wilson RK, Gut IG, Gonder MK, Ryder OA, Hahn BH, Weir BS. 1996. Genetic data analysis II. Sunderland, Massachu- Navarro A, Akey JM, Bertranpetit J, Reich D, Mailund T, setts: Sinauer. Schierup MH, Hvilsom C, Andres AM, Wall JD, Bustamante Wilson ML, Balmforth Z, Cox D, et al. 2009. Pan troglodytes CD, Hammer MF, Eichler EE, Marques-Bonet T. 2013. Great ssp. schweinfurthii. 2009 IUCN red list of threatened species. ape genetic diversity and population history. Nature 499:471– Version 2009.1. Available at: www.iucnredlist.org. Accessed 17 475. March 2010. Pritchard JK, Stephens M, Donnelly P. 2000. Inference of popu- Witzenberger KA, Hochkirch A. 2011. Ex situ conservation lation structure using multilocus genotype data. Genetics genetics: a review of molecular studies on the genetic conse- 155:945–959. quences of captive breeding programmes for endangered ani- Pritchard JK, Wen W, Falush D. 2003. Documentation for mal species. Biodivers Conserv 20:1843–1861. STRUCTURE software: version 2. Available at: http://seq.ege. Wrangham RW, Hagel G, Leighton M, Marshall AJ, Waldau P, fcen.uba.ar/materias/ecomolecular/Material/An%C3%A1lsis%20 Nishida T. 2008. The great ape world heritage species pro- Bayesiano%20Metapoblaciones.%20STRUCTURE/Structure2.1/ ject. In: Stoinski TS, Steklis HD, Mehlman PT, editors. Con- Help%20Files/readme.pdf. Accessed on July 13, 2004. servation in the 21st century: gorillas as a case study. New Raymond M, Rousset F. 1995. GENEPOP (version 1.2): popula- York: Springer Science and Business Media, LLC. p 282– tion genetics software for exact tests and ecumenicism. 296. J Hered 86:248–249. Wright S. 1943. Isolation by distance. Genetics 28:114.

American Journal of Physical Anthropology